The 2007 financial crisis, had a significant impact on Fannie Mae, a government-sponsored enterprise of the U.S. mortgage industry. They provide liquidity to the market by purchasing and guaranteeing mortgages from lenders, enabling more Americans to become homeowners. However, the surge in mortgage defaults during the 2007 crisis put Fannie Mae at risk, leading to government intervention and a bailout. This report analyzes the impact of borrower characteristics on default rates in Fannie Mae’s mortgage portfolio during 2007, comparing it with 2019 to understand trends and changes in borrower behavior over time.

Our analysis began by tracing the changes in mortgage default rates over the years from 2000 to 2023. As seen in Figure 1, the default rates started surging somewhere in 2003 and were at their peak in 2007 at more than 9%. Using an animated chart, we illustrated how default rates have since leveled out, suggesting a market that has learned from the past and is currently on a steadier course with default rates as low as 0.2% in 2019 and declining even further till 2023.

To analyze the reasons for the 2007 financial crisis, we delved into how the occupancy status of borrowers influenced default rates. Figure 2 demonstrates that investors defaulted on their mortgages at a higher rate of nearly 15% compared to those buying homes to live (Principal) or as a second home (Second) whose default rates were almost 9% and 7% respectively in 2007. To analyze this further, we used a stacked bar chart to see how much people from each occupancy status contributed to the total default rate in 2007 and compared it with 2019 to see any similarity. Figure 3 demonstrates that even though investors had the highest default rate in 2007 (nearly 15%), their proportion in the total defaults was 15% while principal owners contributed to 82.5% of the total defaults in 2007. As for 2019, even though the default rate was significantly lower, Principal homeowners account for 94% of the total loan defaults. This signifies that the defaults are majorly from the principal owners.

We next evaluated the purpose of loans for the loan defaults within Fannie Mae’s portfolio. Figure 4 highlights that cash-out refinancing constituted a significant 52% of loan defaults in 2007. This was a risky maneuver that left many homeowners vulnerable when housing prices dropped, leading to an increase in defaults as the loan values surpassed the depreciated home values. In 2019, where the default rates are very significantly lower, the pattern has shifted—purchase loans have emerged as the most defaulted type, accounting for 55% of defaults. This suggests a post-crisis environment characterized by more cautious borrowing and lending.

Another borrower characteristic we analyzed was the debt-to-equity ratio (DTI) using a box plot for 2007 and 2019. Figure 5 shows that the median debt-to-equity ratio was almost as high as 40% in 2007, illustrating a tendency to take on greater debt relative to income. The wider interquartile range (IQR) in 2007 suggests a greater variability in DTI ratios among borrowers. In contrast, the narrower IQR and lower median in 2019 indicate a more uniform distribution of DTI ratios, suggesting a shift towards more conservative borrowing practices post-crisis.

Next, we focused on credit scores for 2007 and 2019, visualizing them using a density plot. Figure 6 illustrates a notable shift in credit score distribution over the years. In 2007, borrowers had credit scores widely spread between 550 to 840, indicating varied creditworthiness during that period. However, in 2019, the distribution shifted significantly, showing a higher concentration of borrowers with credit scores around 800, which is classified as excellent. This shift suggests an improvement in the creditworthiness of borrowers over the years.

The last borrower characteristic we studied was the proportion of total defaults by state in 2007 and 2019. Using an interactive US map chart, we ranked the states based on their default rates. Figure 7 clearly shows that California had the highest defaults in both years, with 13.58% (default count: 612) in 2007 and 18% (default count: 18) in 2019. Florida had the second-highest default rate in 2007, with 515 defaults accounting for 11.42% of the total defaults. Other states with high default rates in 2007 included Illinois, Arizona, and Michigan.

In conclusion, the analysis underscores a transformation in borrower characteristics and default rates in Fannie Mae’s mortgage portfolio from 2007 to 2019. The 2007 crisis was marked by risky borrowing practices, including high debt-to-equity ratios and cash-out refinancing. However, the data reveals a shift towards more conservative borrowing practices post-crisis, with borrowers exhibiting lower debt ratios, higher credit scores, and a preference for purchase loans over cash-out refinancing in 2019. This shift suggests a maturation in the market, as lessons from the crisis have influenced more prudent behavior among borrowers and lenders.

(Word count: 797)


Figure Appendix


Trend of Default Rates Over the Years

Figure 1: Trend of Default Rates Over the Years


Default Rates by Occupancy Status in 2007

Figure 2: Default Rates by Occupancy Status in 2007


Proportion of Total Loan Defaults by Occupancy Status for 2007 and 2019

Figure 3: Proportion of Total Loan Defaults by Occupancy Status for 2007 and 2019


Distribution of Defaulted Loans by Purpose in 2007and 2019

Figure 4: Distribution of Defaulted Loans by Purpose in 2007and 2019


Distribution of Debt-To-Income Ratios for 2007 and 2019

Figure 5: Distribution of Debt-To-Income Ratios for 2007 and 2019


Density Distribution of Borrowers' Credit Scores in 2007and 2019

Figure 6: Density Distribution of Borrowers’ Credit Scores in 2007and 2019


Figure 7: Proportion of Total Defaults by State in 2007 (left) and 2019 (right)